Google Scholar Users Report Badly Malfunctioning Captcha (google.com)

Google's search engine for academic research materials is blocking many users with a malfunctioning captcha screen, according to complaints on a Google help forum. "I'm a doctoral student and a professor, which means I use this extensively. Now I'm blocked from using it at all, even after answering all of the stupid image questions (3 times)," reads a typical complaint.

Heart44 writes: A lot of researchers when using Google Scholar are being asked to prove they are not a robot. You have to find all the rivers (but not the sea or lakes) or all street numbers (but not other numbers) or all the store fronts from nine poor quality images, sometimes more than once and, surprise, you will fail more than two thirds of the time and then just get an error 400 "Malformed request, that's all we know". You are offered an audio challenge but clicking on that simply loads more pictures... Is that the best they can do distinguishing between man and machine?
One post ended by stating succinctly "I'm not a robot, I'm an academic professional, and this process is wasting nontrivial amounts of my time. How do I stop it?"

Comments

EditorDavid did a good job here

[Heart44]

He really improved my submission. He RTFA and made the submission more accessible. Thanks.

Re:This is Google's main problem...

[Richard_J_N]

I completely agree. I had a problem where our new company couldn't send email to Gmail users without always being flagged as spam. We were doing absolutely everything right - and there is no way to get hold of Google. I did finally, 6 months later find a way to reach a person at Google (via a back channel as a customer of a different company), and they confirmed to me: Google act as judge, jury, and executioner, in a secret trial; you can't see the evidence, you don't even know if you've been condemned, and there is no appeal. And they are fine with that.
For what it's worth, the problem was that the previous owners of our IP had got it into a secret blacklist (internal to Google), although we were clean on all of the hundreds of public blacklists I searched. Google are a menace to the public infrastructure. Even AOL behave better!

Re:It finally happened

[SharpFang]

I'm fairly sure the captchas are computer-generated (with Google hoping nobody has as advanced algorithms as they do), because they contain typically computer-related errors.

The "Type the number in" with photo of a building number, shot at an angle, tilted, cropped a little. The number was something like 7375, with the top dash of the first "7" trimmed away by the edge of the picture - but judging by the curve, the tilt, being identical to the second "7", I was confident that was the number.

But no, that answer wasn't accepted. To computer image vision, that's clearly a 1373 and I guess that would be accepted as the captcha answer.

This happens on a more or less regular basis. You shouldn't guess what the actual number is. You should guess what the current, faulty photo makes the number look like. "8" partially obscured by the edge of the building? You'd better type "3", despite the "3" right next to it uses a different shape.